Abstract:The emerging Internet of Things (IoT) era has stimulated many new computation-intensive applications. To support them, mobile edge computing (MEC) is a promising solution that allows users to offload their heavy computing tasks to nearby edge servers. Taking such computation offloading as the service, application service providers (ASPs) can rent resources from mobile network operators for MEC service provisioning. However, it is challenging for ASPs to determine how many resources to rent at different regions and times due to the uncertain user demand. When making an investment strategy, it is crucial to maximize the profit with the consideration on the Quality of Service (QoS), where a joint scheduling on both communication and computing resource under the uncertain demand is needed. To deal with the uncertainty, the probability distribution information is usually employed, which, unfortunately, might be hardly obtainable in practice. Therefore, in this article, we propose a data-driven risk-averse MEC resource investment (DRAI) strategy, where the demand uncertainty issue is particularly addressed. Specifically, we formulate the DRAI strategy into a stochastic optimization problem, which can achieve the expected optimal profit under the QoS guarantee statistically from a risk-averse perspective. To solve it, instead of relying on specific distribution models, we construct an ambiguity set based on the statistical characteristics derived from the historical data that contains all possible distributions, and develop a data-driven distributionally robust solution, aiming at achieving the best strategy under the worst case to make it trustworthy. Simulation results illustrate the effectiveness of the proposed DRAI strategy.

Risk-Averse Investment Strategy for MEC Service Provisioning: A Data-Driven Distributionally Robust Solution

Joint Computation Offloading and Resource Allocation for MEC-Enabled IoT Systems with Imperfect CSI

Data-Driven Optimization for Cooperative Edge Service Provisioning With Demand Uncertainty

Efficient Revenue-Based MEC Server Deployment and Management in Mobile Edge-Cloud Computing

Cost-Effective Traffic Scheduling and Resource Allocation for Edge Service Provisioning

Dynamic Computing Offloading Strategy for Multi-dimensional Resources Based on MEC

DMORA: Decentralized Multi-SP Online Resource Allocation Scheme for Mobile Edge Computing

Dynamic Offloading Strategy for Delay-Sensitive Task in Mobile-Edge Computing Networks

Efficient Risk-Averse Request Allocation for Multi-Access Edge Computing

Resource allocation and network pricing based on double auction in mobile edge computing

Revenue-Maximized Offloading Decision And Fine-Grained Resource Allocation In Edge Network

Distributionally Robust Optimization for Computation Offloading in Aerial Access Networks

Adaptive Request Scheduling and Service Caching for MEC-Assisted IoT Networks: an Online Learning Approach

Dynamic User Allocation in Stochastic Mobile Edge Computing Systems

An Efficient Computation Offloading and Resource Allocation Algorithm in RIS Empowered MEC

PDMA: Probabilistic Service Migration Approach for Delay-aware and Mobility-aware Mobile Edge Computing

Age-of-Information-Based Computation Offloading and Transmission Scheduling in Mobile-Edge-Computing-Enabled IoT Networks

Cost-Effective and Robust Service Provisioning in Multi-Access Edge Computing

Platform Profit Maximization in D2D Collaboration based Multi-Access Edge Computing

Towards Effective Resource Procurement in MEC: a Resource Re-selling Framework

Optimal Resource Reservation Scheme for Maximizing Profit of Service Providers in Edge Computing Federation